Question 1010
An application expects a model to use a 100,000-token context, but the selected deployment has a lower configured cap despite model documentation mentioning a larger maximum. What must be verified?
Explanation
Correct Answer
Advertised model capabilities can differ by deployment route and configuration; validate the actual served limit.
Incorrect Answers
Subscription names do not determine token constraints.
Source comments do not change model limits.
Training batch sizes are not chat context windows.